基于软间隔支持向量机的装配式建筑构件质量控制研究

Research on Quality Control of Prefabricated Building Components Based on Soft Spaced Support Vector Machine

相比较传统现浇结构,装配式建筑因其构件拆分的特点,需要满足住宅结构方案和深化设计的要求,从全生命周期角度分析装配式建筑深化设计和构件拆分、预制生产和吊装运输、施工现场安装、运维等各个环节质量缺陷。利用支持向量机模型对装配式建筑质量的量化特征,以及量化特征的平均值、标准差和最坏值进行模式识别,划分最佳超平面。利用核函数将线性不可分的向量转换成高维度特征空间中线性可分向量,得到对于装配式建筑建造过程中质量的准确识别和判断。

Compared with traditional cast-in-situ structures,prefabricated buildings need to meet the requirements of residential structure schemes and deeper design due to the characteristics of component splitting.From the perspective of whole life cycle,this paper analyzes the quality defects from the aspects of deepening design and component splitting,prefabrication production and hoisting transportation,construction site installation,operation and maintenance,etc.Then,uses the support vector machine model for pattern recognition of the quantified features of prefabricated building quality,as well as the average,standard deviation,and worst value of the quantified features to divide the best hyperplane.Finally,uses the kernel function to convert the linear inseparable vector into the linearly separable vector in the high-dimensional feature space to obtain accurate identification and judgment of the quality during the construction of prefabricated building.